A rail transit simulation system for multi-modal energy-efficient routing applications

Abstract

The paper develops a continuous rail transit simulator (RailSIM) intended for multi-modal energy-efficient routing applications. RailSIM integrates sophisticated train dynamics and energy models to replicate train motion and energy consumption behavior, respectively. The simulator is calibrated using an off-line optimization procedure to match preprogramed railway schedules by optimizing three model parameters, namely; the segment target speed, the average deceleration level, and the brake force adjustment factor. The objective of the calibration procedure is to match the simulated and actual average running speed for each station-to-station pair. Upon calibration, RailSIM is applied to the Greater Los Angeles area and validated at both the instantaneous and aggregated levels. Results demonstrate that RailSIM is able to produce realistic train dynamics and energy consumption estimates producing a comfortable ride while simultaneously matching the railway schedule. RailSIM is also demonstrated to capture the impact of track gradient on energy outputs. The results also indicate that a perfect match to empirical energy estimates is achieved at an average grade of 1.8%, which is a reasonable approximation of the average track gradient of the testing area. The sensitivity of RailSIM to some of the metro rail parameters is also discussed to account for its applicability to rail transit systems in other cities. Finally, a pilot test of RailSIM implementation in the higher-level multi-modal eco-routing system is performed to demonstrate RailSIM’s feasibility of supporting energy-efficient travel.

Keywords:

• Eco-routing
• energy consumption
• multi-modal
• rail transit
• simulation
• smart mobility

Acknowledgments

The authors would like to thank Dr. Matthew Klenk and Dr. Lei Lin for providing the LA data used to validate the model results.

Notes

1 The brake force adjustment factor is applied to determine the brake level.

2 The General Transit Feed Specification (GTFS) defines a common format for public transportation schedules and associated geographic information.

3 F_s is the starting tractive effort required to move a train from a complete stop; it equals 0 when the train is moving.

4 The target speed should be greater than average speed to guarantee that the simulated average speed is not less than the real value.

5 The orange line is a metro bus line.

6 Some of the rolling stock parameters are considered static inputs. These include the drag coefficient and maximum hotel load, while some of them are dynamic such as the number of cars per train and weight per railcar, which depend on the rail line, moving direction, and time of operation.

7 The real-world average speeds for a few track segments in the LA rail network are greater than the train maximum allowed running speed U_max, which is absolutely not correct. To guarantee a complete stop for the train at the downstream station of each of these segments, we set a high brake level which results in extremely large deceleration (e.g., –5.2$m/s^2$). Nonetheless, this setting guarantees the incorrect real-world average speeds not to affect the overall performance of the simulation model and enables the model error-tolerant. The deceleration of –5.2 $m/s^2$ only occurs on those segments with incorrect real-world average speeds; for other segments, deceleration is no greater than 2.0 $m/s^2,$ as illustrated in Figure 5.

8 A significant number of track segments are underground in the L.A. metro rail system.

Additional information

Funding

This work was jointly funded by the Department of Energy through the Advanced Research Projects Agency-Energy (ARPA-E) TRANSNET Program under award number DE-AR0000612 and by the University Equity and Mobility Center (UMEC) under award number 69A3551747123.